Across five US surveys, more people view AI negatively than positively, even though the questions differ. Our World in Data’s analysis finds that 52% of respondents in a Pew survey were more concerned than excited about AI, against 9% who were more excited.
Watch Desk analysis
What happened
Max Roser of Our World in Data compared five recent surveys using probability-based samples, rather than relying on opt-in online polls. The surveys were weighted to match the US adult population, and their full questions and results are published.
The precise numbers vary with the wording. In Pew’s survey, 37% chose a middle response when it was offered. In both Pew’s and Gallup’s surveys, 9% expressed a positive view. An Ipsos survey found 17% considered AI extremely bad for society, while 2% considered it extremely good.
Why it matters
There is no single polling question that captures the public’s view of AI. But the broad pattern in these surveys is not a close call: negative views outnumber positive ones. That matters for companies and policymakers making choices about AI’s deployment, oversight and public case for support.
The findings are not a referendum on every application of AI. The article notes that one Ipsos question about AI’s effect on people personally produced a more even split: 50% said it was good for them personally, and 49% said bad. The question asked changes the answer, which is rather the point of doing more than one poll.
Our read
This is a useful snapshot of public sentiment, not proof that every American is wary or that opinion cannot change. The strongest takeaway is the consistency across differently worded surveys, alongside a reminder to check what a poll actually asked before turning its headline into a claim about “the public”.
What to watch
- Whether future probability-based surveys find the same gap between concern and enthusiasm.
- How views differ between AI’s effects on society and its effects on people personally.
- Whether AI companies respond to public concerns with changes people can see, rather than another round of reassurance.
Discussion spark: What would do more to earn public trust in AI: clearer evidence about its benefits, stronger rules on how it is used, or a greater say for the people affected?
Sources and evidence
- Source update (16 November 2026, 00:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.